An ensemble machine learning approach for classification tasks using feature generation

Connection Science, Journal Year: 2023, Volume and Issue: 35(1)

Published: July 11, 2023

Although machine learning classifiers have been successfully used in the medical and engineering fields, there is still room for improving predictive accuracy of model classification. The higher classifier, better suggestions can be provided decision makers. Therefore, this study, we propose an ensemble approach, called Feature generation-based Ensemble Support Vector Machine (FESVM), classification tasks. We first apply feature selection technique to select most related features. Next, introduce strategy aggregate multiple base estimators final prediction using meta-classifier SVM. During stage, use probabilities obtained from classifier generate new After that, generated features are added original data set form a set. Finally, utilised train SVM obtain results. For example, binary task, each has two (p one class 1−p other class). In case, combination based on these classifiers. One sum p as 1, 2. These then same way, our generation method easily extended multi-class task generating features, where number depends classes. Those (first layer) This input second layer (meta-classifier) model. Experiments 20 sets show that proposed FESVM best performance compared under comparison. addition, than stacking Statistical results Wilcoxon–Holm also confirms significantly outperform models. indicate useful tool tasks, especially multi-classification

Language: Английский

---

A Novel Fuzzy Feature Generation Approach for Happiness Prediction

IEEE Transactions on Emerging Topics in Computational Intelligence, Journal Year: 2024, Volume and Issue: 8(2), P. 1595 - 1608

Published: Jan. 24, 2024

Happiness refers to an emotional state of well-being and contentment. Accurate prediction happiness is important for people in promoting a healthy lifestyle, helping reduce stress, enhancing humans' immune system. In this paper, we propose novel fuzzy feature generation approach prediction. We design weighted operation based on the IF-THEN rules generate feature. This generated (new information) added model training achieve more accurate model. addition, considering high interpretability rules, it can improve process generation. Experimental results show that with use proposed approach, performance used machine learning models be improved prediction, outperforming state-of-the-art models. Among all models, FF-CatBoost has best terms accuracy (62.75%) F1-score (66.63%). Results other data sets also confirm effectiveness our approach. The statistical from Wilcoxon rank-sum test further significantly accuracy. With its excellent useful tool help know about their status live happier life.

Language: Английский

---

A Feature Importance-Based Multi-Layer CatBoost for Student Performance Prediction

IEEE Transactions on Knowledge and Data Engineering, Journal Year: 2024, Volume and Issue: 36(11), P. 5495 - 5507

Published: May 3, 2024

Student performance prediction is vital for identifying at-risk students and providing support to help them succeed academically. In this paper, we propose a feature importance-based multi-layer CatBoost approach predict the students' grade in period exam. The idea construct structure with increasingly important features layer by layer. Specifically, importance are first calculated sorted ascending order. each layer, least accumulated until reaching given threshold. Then, these selected used training CatBoost. Next, trained utilized generate that adds set their within After that, all train next This process repeated used. results show proposed model has best performance. Moreover, statistical test conducted based on 20-runs of experiments validates significant superiority our over compared models demonstrates efficacy enhancing model. indicates can decision makers educational quality.

Language: Английский

---

Sampling-Based Machine Learning Models for Intrusion Detection in Imbalanced Dataset

Electronics, Journal Year: 2024, Volume and Issue: 13(10), P. 1878 - 1878

Published: May 11, 2024

Cybersecurity is one of the important considerations when adopting IoT devices in smart applications. Even though a huge volume data available, related to attacks are generally significantly smaller proportion. Although machine learning models have been successfully applied for detecting security on applications, their performance affected by problem such imbalance. In this case, prediction model preferable majority class, while predicting minority class poor. To address problems, we apply two oversampling techniques and undersampling balance different categories. verify performance, five models, namely decision tree, multi-layer perception, random forest, XGBoost, CatBoost, used experiments based grid search with 10-fold cross-validation parameter tuning. The results show that both can improve used. Based results, XGBoost SMOTE has best terms accuracy at 75%, weighted average precision 82%, recall F1 score 78%, Matthews correlation coefficient 72%. This indicates technique effective multi-attack under imbalance scenario.

Language: Английский

---

Predicting secondary school student performance using a double particle swarm optimization-based categorical boosting model

Engineering Applications of Artificial Intelligence, Journal Year: 2023, Volume and Issue: 124, P. 106649 - 106649

Published: June 24, 2023

Language: Английский

---

A comprehensive wind power prediction system based on correct multiscale clustering ensemble, similarity matching, and improved whale optimization algorithm–A case study in China

Renewable Energy, Journal Year: 2025, Volume and Issue: unknown, P. 122529 - 122529

Published: Feb. 1, 2025

Language: Английский

---

Quantum encoding whale optimization algorithm for global optimization and adaptive infinite impulse response system identification

Artificial Intelligence Review, Journal Year: 2025, Volume and Issue: 58(5)

Published: March 9, 2025

Language: Английский

---

A Hybrid Nonlinear Whale Optimization Algorithm with Sine Cosine for Global Optimization

Biomimetics, Journal Year: 2024, Volume and Issue: 9(10), P. 602 - 602

Published: Oct. 7, 2024

The whale optimization algorithm (WOA) is constructed on a whale's bubble-net scavenging pattern and emulates encompassing prey, devouring stochastic capturing for prey to establish the global optimal values. Nevertheless, WOA has multiple deficiencies, such as restricted precision, sluggish convergence acceleration, insufficient population variety, easy premature convergence, operational efficiency. sine cosine (SCA) oscillation attributes of coefficients in mathematics methodology. SCA upgrades amplifies search region, accelerates international investigation regional extraction. Therefore, hybrid nonlinear with (SCWOA) emphasized estimate benchmark functions engineering designs, ultimate intention investigate reasonable solutions. Compared other algorithms, BA, CapSA, MFO, MVO, SAO, MDWA, WOA, SCWOA exemplifies superior effectiveness greater computation profitability. experimental results emphasize that not only integrates extraction avoid realize most appropriate solution but also exhibits superiority practicability locate precision faster speed.

Language: Английский

---

Integrating the Opposition Nelder–Mead Algorithm into the Selection Phase of the Genetic Algorithm for Enhanced Optimization

Applied System Innovation, Journal Year: 2023, Volume and Issue: 6(5), P. 80 - 80

Published: Sept. 4, 2023

In this paper, we propose a novel methodology that combines the opposition Nelder–Mead algorithm and selection phase of genetic algorithm. This integration aims to enhance performance overall To evaluate effectiveness our methodology, conducted comprehensive comparative study involving 11 state-of-the-art algorithms renowned for their exceptional in 2022 IEEE Congress on Evolutionary Computation (CEC 2022). Following rigorous analysis, which included Friedman test subsequent Dunn’s post hoc test, demonstrated outstanding performance. fact, exhibited equal or superior compared other majority cases examined. These results highlight competitiveness proposed approach, showcasing its potential achieve solving optimization problems.

Language: Английский

---

VWFTS-PSO: a novel method for time series forecasting using variational weighted fuzzy time series and particle swarm optimization

International Journal of General Systems, Journal Year: 2024, Volume and Issue: unknown, P. 1 - 20

Published: Sept. 23, 2024

Language: Английский

---